1.

Introduction

Minimally invasive surgery aims at minimizing the surgical trauma and reducing the patient’s recovery time.¹ Lasers play an increasingly important role in this new approach to medical care, because they allow the noncontact cutting and removal of a wide variety of living tissues with extreme accuracy and reduced mechanical and thermal impacts, and they decrease the risk of collateral damage to peripheral structures such as nerves and blood vessels. Moreover, laser beams can be transported at distance by optical fiber and are essentially inertialess, facilitating remote operation by manipulators and robots. Lasers are also fully compatible with existing laparoscopic techniques and robotic surgery.² As a result, they are being actively investigated as a viable alternative to mechanical tools for delicate surgical applications at complicated anatomical locations, in particular for orthopaedic and cranio-maxillo-facial surgery.³

Bone is the main tissue in the skeletal system. It contains about 70% hydroxyapatite and 30% proteins (mainly collagen) and water. Its constitution is similar to that of dentin, but the two tissues present large differences in structure and functionality. Teeth consist mainly of intertubular dentin formed by a network of collagen fibers reinforced by hydroxyapatite nanocrystals, traversed by hollow cylindrical formations consisting of crystalline hydroxyapatite (peritubular dentin). Cells and blood vessels concentrate in the tooth central region, designated by pulp. Bone presents a much more complex structure; each bone includes a compact outer layer of cortical bone with porosity in the range of 5% to 30%, where hydroxyapatite nanocrystals surround the blood vessels, and an inner core of extremely porous cancellous bone (porosity 30% to 90%), consisting of a complex network of hydroxyapatite-rich plates and beams that limit cavities. These cavities are occupied by blood vessels and living cells.⁴ In general, bone presents a larger content of living cells than dentin, enabling the tissue to regenerate itself continuously.

Lasers have been investigated for hard-tissue cutting and removal since 1964. Ruby,^5,6 ${\mathrm{CO}}_2$,^7,8 Nd:YAG,⁹ ${\mathrm{Ar}}^{+}$,¹⁰ Ho:YAG,¹¹ and alexandrite lasers¹² have all been tested for dentin removal without entirely fulfilling the existing expectations, in particular due to their low ablation rate and excessive heating of the tooth, potentially leading to carbonization, thermomechanical cracking, and, in extreme situations, necrosis.¹² These drawbacks can be overcome by using lasers favoring athermal instead of thermal ablation. Studies performed by our team¹³ showed that dentin processed with 30 ns pulse duration, 248 nm wavelength excimer laser radiation presents no chemical or structural degradation. With these lasers, ablation is predominantly photochemical and, globally, heating is negligible. The reason for this behavior is that the ablation of intertubular dentin, the main constituent of the tooth bulk, is controlled by the photochemical decomposition of collagen, thus minimizing the thermal impact of the laser treatment on the material.^14,15 Unfortunately, the ablation rates achieved with these lasers are too low for practical applications, and their intense UV beams are a potential health hazard.¹⁶ Er:YAG and Er,Cr:YSGG lasers (wavelengths of 2.94 μm and 2.78 μm, respectively) allow achieving higher ablation rates for dentin and are now extensively used in dentistry.^17–19 The radiation wavelengths of these lasers are within strong absorption bands of water. As a result, radiation is efficiently absorbed by the water present in dentin. The water is suddenly evaporated, leading to a pressure pulse that mechanically disrupts the tissue.²⁰ The resulting removal rates are high, and thermal damage is low, as compared to other pulsed infrared lasers, but there are indications that, for certain parameter ranges, the mechanical shock may lead to subsuperficial cracking of dentin.^20,21

Laser ablation of bone was less studied than the ablation of dentin, but since the constitution of the two tissues is similar, identical ablation behavior can be expected. ${\mathrm{CO}}_2$ lasers originated severe charring in bone cutting experiments.²² On the contrary, Er:YAG and Er,Cr:YSGG lasers were successfully tested for orthopaedic, maxillofacial, and periodontal surgery,^17–19 but thermal necrosis of the tissue was observed in some conditions.²³ Arguably, these detrimental effects can be overcome by using lasers with subpicosecond pulse duration,²⁴ because, at the high radiation intensities achieved with these lasers (larger than ${10}^{11}\mathrm{W}/{\mathrm{cm}}^2$), multiphoton ionization becomes probable. The resulting free electrons are accelerated by the incoming radiation and collide with other atoms, initiating an avalanche effect that leads to the formation of a plasma and, eventually, to ablation.²⁴ Plasma-induced ablation of dentin allows relatively high ablation rates without melting or cracking and with negligible modification of the underlying tissue constitution.²⁵ Wieger, Zoppel, and Wintner²⁶ compared the ablation of cortical and cancellous bovine bone by Yb picosecond pulse duration and Er nanosecond pulse duration lasers and found numerous advantages in the first type of laser, namely high removal rates and absence of collateral damage such as melting, carbonization, and cracking. On the other hand, Leucht et al. observed²⁷ that the regeneration of a corticotomy performed with a Ti:Sapphire laser with a 1 ps pulse duration is faster than when the procedure was carried out by conventional saw cutting. The authors explained this difference by the absence of denaturation of the collagen of bone, which reduces the probability of inflammatory reactions.

Ultrafast lasers are promising tools for orthopaedic surgery, but we are still far from understanding femtosecond laser bone ablation mechanisms, and no detailed studies of the chemical and structural modifications induced in bone by the laser treatment exist up to the moment. The aim of the present work was to study the topography, structure, and constitution of bovine femoral bone surfaces treated with femtosecond infrared laser radiation. Samples of cortical bone were harvested and submitted to treatment with 500 fs, 1030 nm wavelength laser pulses. The surface topography of the treated surfaces was studied by scanning electron microscopy (SEM), and their constitution was investigated by glancing incidence x-ray diffraction²⁸ (GI-XRD), x-ray photoelectron spectroscopy²⁹ (XPS), Fourier transform infrared spectroscopy (FTIR), and micro-Raman spectroscopy³⁰ techniques, which were selected because they provide complementary information on material surface composition and structure. Acoustic emission measurements³¹ were performed to detect the onset of the ablation process.

2.

Materials and Methods

3.

Results

3.1.

Ablation Threshold

The ablation threshold was calculated using the $D^2$-method³² described in Sec. 2. The variation of the square diameter of the craters as a function of laser fluence for 5, 10, 15, and 100 laser pulses is represented in the semilogarithmic plots of Fig. 1, together with linear fittings to the experimental data. The corresponding ablation thresholds, determined by extrapolating $D^2$ to zero, are $0.59\pm 0.08$, $0.48\pm 0.06$, $0.48\pm 0.06$, and $0.32\pm 0.04\mathrm{J}/{\mathrm{cm}}^2$, respectively ($R^2:0.95\%$, 0.99%, 0.98%, and 0.91%). The incubation coefficient and the ablation threshold for a single pulse were determined from the plot of the accumulated fluence [$N·{\varphi }_{\mathrm{th}}$ ($N$)] as a function of the number of pulses depicted in Fig. 2 ($R^2=0.9998$) and are $0.81\pm 0.01$ and $0.79\pm 0.04\mathrm{J}/{\mathrm{cm}}^2$, respectively.

Fig. 1

Semilogarithmic plot of $D^2$ as a function of fluence. The lines are linear fitting of the experimental data.

Fig. 2

Accumulated fluence as a function of the number of pulses. The solid line represents the fitting according to Eq. (5).

A crater created by 30 laser pulses and 340 μJ pulse energy (average fluence $1.09\mathrm{J}/{\mathrm{cm}}^2$) and the corresponding depth profile are presented in Fig. 3. The ablated volume increases with the number of pulses up to a saturation value that increases with fluence (Fig. 4). The average saturation values, determined by averaging the ablation volume values for 200, 300, 400, and 500 laser pulses, are 1.8, 3.9, and $5.3\times {10}^5{\mu \text{m}}^3$, respectively.

Fig. 3

Crater produced with a fluence of $1.09\mathrm{J}/{\mathrm{cm}}^2$ and 30 pulses at 10 Hz. (a) SEM image; (b) 2D profile.

Fig. 4

Variation of the ablated volume as a function of the number of pulses (from 20 to 500). The curves result from a fitting to the experimental data and are a guide to the eye.

The variation of the maximum acoustic emission signal as a function of fluence is plotted in Fig. 5. Two variation regimes can be identified, both characterized by a logarithmic dependence of the acoustic signal on fluence. For fluences lower than $0.6\mathrm{J}/{\mathrm{cm}}^2$, the maximum acoustic emission intensity depends only weakly on fluence, the experimental data being adequately fitted by the equation $y=-473+258\times \ln (x+5.94)$ ($R^2=0.99$). For higher fluences, this dependence is more accentuated. The experimental data was fitted by the function $y=47+78\times \ln (x+0.02)$ with $R^2=0.97$.

Fig. 5

Variation of the acoustic emission during the 10th laser pulse as a function of the radiation fluence.

3.2.

Surface Morphology

The topography of a cortical bone surface created by ablation with an average fluence of $1.09\mathrm{J}/{\mathrm{cm}}^2$ is depicted in the SEM micrograph of Fig. 6, which corresponds to a magnification of the crater in Fig. 3(a). The surface is flat, with an irregular and ragged appearance. The bone canaliculi are exposed, and the bone structure is preserved. No traces of resolidified material or cracks are observed. The surface is also free of major redeposited ablation debris or spallation ejecta created by mechanical disruption of the material by expanding gases due to water vaporization or tissue decomposition.

Fig. 6

Surface produced with a fluence of $1.09\mathrm{J}/{\mathrm{cm}}^2$ and 30 pulses at 10 Hz, corresponding to a magnification of the inside of the crater depicted in Fig. 3(a).

3.3.

Surface Composition

3.3.1.

Fourier transform infrared spectroscopy

The ATR-FTIR spectra of cortical bone in the untreated condition and after laser treatment with average fluences of 0.55, 0.82, and $2.18\mathrm{J}/{\mathrm{cm}}^2$ are represented in Fig. 7. The spectrum of the untreated sample has a higher signal-to-noise ratio than the laser-treated samples’ spectra, because the sample was polished and presented higher reflectivity. The phosphate and carbonate absorption bands in these spectra are associated with the mineral phase of bone, a calcium-deficient hydroxyapatite. The most intense bands, associated with ${{\mathrm{PO}}_4}^{3-}$, appear at 1010 and $956{\mathrm{cm}}^{-1}$ and correspond to the ${\nu }_3$ antisymmetric PO stretching mode and the ${\nu }_1$ symmetric stretching mode, respectively.³⁵ The absorption bands at 1415 and $1450{\mathrm{cm}}^{-1}$ can be attributed to the ${\nu }_3$ mode of ${{\mathrm{CO}}_3}^{-2}$ substituted in B-type ${{\mathrm{PO}}_4}^{3-}$ and A-type ${\mathrm{OH}}^{-}$ anionic sites, respectively.^13,35 The band at $873{\mathrm{cm}}^{-1}$ is related to ${{\mathrm{CO}}_3}^{-2}$ ${\nu }_2$ mode.³⁵ The organic constituents of the material are revealed by the bands of collagen, the main organic component of bone, between 1650 and $1200{\mathrm{cm}}^{-1}$. The bands at 1680 to $1600{\mathrm{cm}}^{-1}$, 1580 to $1510{\mathrm{cm}}^{-1}$, and $1240{\mathrm{cm}}^{-1}$ are related to the amide I ($\mathrm{C}=\mathrm{O}$ bond stretching), amide II (C-N bond stretching and N-H deformation modes), and amide III groups of the collagen molecule, respectively. The broad absorption band centered at $3320{\mathrm{cm}}^{-1}$ can be attributed to amide A group (N-H asymmetric stretching mode) of collagen, and the band at 3000 to $2900{\mathrm{cm}}^{-1}$ corresponds to the stretching of the ${\mathrm{CH}}_2$ chain bonds.³⁶ Laser ablation leads to a decrease in the relative amplitude of the amide bands associated with collagen. The relative amplitude of the other IR absorption peaks, namely those associated with the mineral constituent of bone, is not significantly affected by the laser treatment.

Fig. 7

FTIR spectra of treated and untreated samples.

3.3.2.

Micro-Raman spectroscopy

The micro-Raman spectrum of a sample treated with $0.82\mathrm{J}/{\mathrm{cm}}^2$ is depicted in Fig. 8, together with the spectrum of an untreated sample. All the Raman bands observed in the laser-treated sample spectrum can be unequivocally assigned to bone’s constituents, showing that no new phases formed due to the laser treatment. No peaks corresponding to amorphous carbon are observed (they should exist³⁷ at ca. 1580 and $1350{\mathrm{cm}}^{-1}$), showing that no carbonization occurred. The spectra of Fig. 8(a) present the main phosphate ${\nu }_1$ vibration band at ca. $960{\mathrm{cm}}^{-1}$ (the higher amplitude band) and the two main collagen bands, the amide I and III at ca. 1676 and $1247{\mathrm{cm}}^{-1}$, the bending vibration band of the ${{\mathrm{HPO}}_4}^{2-}$ ion at ca. $1453{\mathrm{cm}}^{-1}$, and the ${\nu }_1$ vibration band of ${{\mathrm{CO}}_3}^{2-}$ at ca. $1070{\mathrm{cm}}^{-1}$ (Ref. 38). They are similar for both sample conditions. Figure 8(b) shows an expanded view of the overlapped CH vibration bands of the lipidic and proteic components of the tissue. The differences between the Raman spectra of untreated and treated samples are very small. The spectrum of the untreated sample presents a better signal-to-noise ratio in the collagen bands and a higher relative intensity of the ${\nu }_3$ vibration band of the phosphate group at ca. $1004{\mathrm{cm}}^{-1}$ as compared to the laser-treated samples’ spectrum, probably due to the different surface roughness of the samples. The most relevant difference between the spectra of the treated and untreated samples is the band at ca. $855{\mathrm{cm}}^{-1}$, corresponding to the stretching mode of the aromatic $\mathrm{C}─\mathrm{C}─\mathrm{H}$ bonds,³⁸ which is present in the spectrum of the untreated sample and absent from the spectrum of the treated sample.

Fig. 8

Micro-Raman spectra of (a) the region between 400 and $1800{\mathrm{cm}}^{-1}$ from an untreated bone tissue sample (i) and treated bone tissue sample (ii); (b) the region between 2700 and $3200{\mathrm{cm}}^{-1}$ from an untreated bone tissue sample (i) and treated bone tissue sample (ii).

3.3.3.

X-ray diffraction

The glancing-incidence x-ray diffractograms of cortical bone before and after laser treatment are shown in Fig. 9, together with a diffractogram of synthetic hydroxyapatite presented as a reference. All the diffraction peaks observed in bone diffractograms can be indexed as nonstoichiometric hydroxyapatite,³⁹ according to JCPDS card number 09-0432. The large width of the hydroxyapatite peaks in bone is consistent with the partially amorphous nature of the phase. The diffractograms of the samples treated with 0.82 and $1.09\mathrm{J}/{\mathrm{cm}}^2$ present a prominent narrow peak at 32.8 deg that can be indexed as the $(\begin{array}{ccc}3& 0& 0\end{array})$ peak of hydroxyapatite and that is not observed in the diffractogram of the untreated sample.

Fig. 9

X-ray diffractograms of treated and untreated bone samples and synthetic hydroxyapatite.

3.3.4.

X-ray photoelectron spectroscopy

The regions of the XPS spectra retained for analysis are those related to hydroxyapatite, i.e., the peaks associated to Ca 2p, O 1s, and P 2p (Fig. 10). The spectral regions corresponding to magnesium, an element that partially replaces calcium in the hydroxyapatite crystallographic structure of natural hard tissues,⁴⁰ are not described, because they did not provide relevant information, while the C 1s and nitrogen regions—originating from the collagen molecules—were considered only for energy calibration purposes, because carbon contamination precludes the quantification of these elements. The Ca 2p region of the spectra exhibits a single doublet both for the laser-treated and untreated samples. The main component of the doublet ($\mathrm{Ca}2{\mathrm{p}}_{3/2}$) is centered at a binding energy of $347.4\pm 0.1\mathrm{eV}$, while its second component ($\mathrm{Ca}2{\mathrm{p}}_{1/2}$) has an orbit-spin shift of $3.6\pm 0.1\mathrm{eV}$, as shown in Fig. 10(a). The P 2p peak consists of a single doublet, as well, with a separation of $0.9\pm 0.1\mathrm{eV}$, the component $\mathrm{P}2{\mathrm{p}}_{3/2}$ being centered at $133.2\pm 0.1\mathrm{eV}$, as shown in Fig. 10(b) and in agreement with the results of Sivakumar et al.⁴¹ The O 1s region was fitted with a main component at $531.1\pm 0.1\mathrm{eV}$, as shown in Fig. 10(c). This can be assigned to oxygen in the hydroxyapatite phosphate group and to oxygen from carbonyl groups. The second component is about 20% of the main one in the untreated region of the sample and decreases to less than 10% for the treated region. The main results of the quantitative analysis of the XPS spectra are presented in Table 1.

Fig. 10

Relevant regions of the XPS spectra of untreated and laser-treated cortical bone samples: (a) Ca 2p; (b) P 2p; (c) O 1.

Table 1

XPS atomic percentages and atomic ratios in the treated and untreated regions.

	Treated	Untreated
Ca	14.9	10.2
Mg	0.9	n.m.
P	9.7	7.0
${\mathrm{O}}_{531.1}$	42.3	33.1
${\mathrm{O}}_{532.7}$	3.8	6.3
N	5.0	7.7
C	23.4	35.6
$\mathrm{Ca}/\mathrm{P}$	1.54	1.46
${\mathrm{O}}_{531.1}/\mathrm{P}$	4.4	4.7
$(\mathrm{Ca}+\mathrm{Mg})/\mathrm{P}$	1.64

4.

Discussion

The results described in Sec. 3 show that the threshold for the ablation of bovine cortical bone with 500 fs duration pulses of 1030 nm laser radiation decreases when the number of laser pulses increases (Table 2). This variation can be explained by the creation of defects at fluences lower than the single-pulse ablation threshold, which facilitates ablation by multiphoton excitation during subsequent laser pulses.⁴² For infrared photons with energy lower than the energy gap of dielectrics, radiation absorption in perfect crystals can be achieved only by multiphoton absorption. However, it is often verified that laser irradiation at fluences lower than the ablation threshold, despite being insufficient to promote direct ablation during the first laser pulses, may act to create defects corresponding to energy levels within the gap that enhance radiation absorption by the material and facilitate ablation during subsequent laser pulses. Assuming that this mechanism was occurring, a value for the incubation coefficient of $0.81\pm 0.01$ was calculated using the method proposed by Jee, Becker, and Walser.³³ This value is in good agreement with the one determined by Nicolodelli, Lizarelli, and Bagnato⁴³ for bovine femur bone treated with a Ti:Sapphire femtosecond with 70 fs pulse duration and 801 nm central emission wavelength ($S=0.788\pm 0.004$).

Table 2

Ablation threshold for different numbers of laser pulses calculated using the D2-method.

The single-pulse ablation threshold of cortical bone in the present experimental conditions is $0.79\pm 0.04\mathrm{J}/{\mathrm{cm}}^2$. This value is compared to those found in the literature in Table 3, which also includes the main processing parameters used by the different authors. The present value is in good agreement with the one determined by Wieger, Zoppel, and Wintner²⁶ for bovine cortical bone and by Girard et al.⁴⁴ for porcine cortical bone ($0.69\pm 0.08\mathrm{J}/{\mathrm{cm}}^2$) and higher than the ablation threshold found by Poli⁴⁵ for rabbit cortical bone ($0.54\mathrm{J}/{\mathrm{cm}}^2$). The ablation thresholds determined by Emigh et al.⁴⁶ for unpolished samples of porcine cortical bone and by Altman⁴⁷ for bovine femoral bone are much higher (3.29 and $2.6\mathrm{J}/{\mathrm{cm}}^2$, respectively). Taking into consideration that dentin has a constitution similar to that of bone, it is instructive to compare these values with those found for dental hard tissues, and the corresponding values were included in the table. Neev et al.⁴⁸ found 0.5 and $0.7\mathrm{J}/{\mathrm{cm}}^2$ for the ablation thresholds of dentin and enamel, respectively, using a laser with 350 fs pulse duration and 1050 nm wavelength, while Alves, Oliveira, and Vilar²⁵ found $0.6\pm 0.2\mathrm{J}/{\mathrm{cm}}^2$ for the ablation threshold of dentin in experimental conditions similar to those used in the present work. The lower ablation threshold value found by Poli⁴⁵ can probably be explained by the extremely low pulse duration of the laser used by this author (51 fs). The other values fall into two classes, the first with values around $0.8\mathrm{J}/{\mathrm{cm}}^2$ and the second with values between 2.6 and $3.3\mathrm{J}/{\mathrm{cm}}^2$. Emigh et al.⁴⁶ explained the large ablation threshold found in their work by the fact that the samples used were not polished, so they presented lower reflectivity. However, this explanation does not apply to the results of Altman,⁴⁷ because this author used polished samples. A more plausible explanation results from the consideration of the spectral dependence of the optical properties of bone. The absorption coefficient of cortical bone for 800 and 1000 nm is $0.11\pm 0.02$ and $0.22\pm 0.03{\mathrm{cm}}^{-1}$, respectively.⁴⁹ As a result, the radiation absorption depth is smaller for 1000 nm than for 800 nm, leading to higher volumetric energy density. This difference is further enhanced when the radiation scattering is considered, as well.²⁴

Table 3

Hard tissues’ ablation threshold values obtained with ultrashort pulse duration lasers and their corresponding radiation wavelength.

The ablated volume increases with number of pulses up to a saturation value (Fig. 4), and this saturation can be attributed to the progressive increase of the ablation crater depth and surface curvature with the increasing number of pulses, which leads to a continuous decrease of the fluence. The fact that no physicochemical or topographic alterations of the surface resulting from the laser treatment were observed can also explain this saturation effect.

The acoustic emission analysis revealed two regimes of variation of the maximum acoustic emission intensity with fluence. Both regimes are characterized by a logarithmic dependence of the acoustic signal on fluence, but for fluences lower than $0.6\mathrm{J}/{\mathrm{cm}}^2$, the maximum acoustic emission intensity shows a weak logarithmic dependence on fluence, while for higher fluences, this dependence is more accentuated. Nolte et al.⁵⁰ observed a similar behavior in the subpicosecond laser ablation of some metals and proposed a theoretical model explaining it. According to the authors, in the first regime, corresponding to the low-fluence range, ablation is controlled by the optical penetration depth of the material, while in the second one, ablation is determined by the effective heat penetration depth. However, in the present case, the low-fluence regime extends to fluences lower than the material ablation threshold for a similar number of pulses ($0.48\pm 0.06\mathrm{J}/{\mathrm{cm}}^2$), so this regime cannot be associated with an ablative process. Instead, it may correspond to subablative thermal processes, such as the intense evaporation and desorption of volatile components, leading to the formation of a rapidly expanding gas plume^24,51 or to thermoelastic phenomena.^51,52 This interpretation is corroborated by the fact that the fluence found by extrapolating the curve fitted to the high-fluence regime data in Fig. 5 ($0.53\mathrm{J}/{\mathrm{cm}}^2$) is similar to the 10 pulse ablation threshold calculated by the $D^2$-method ($0.48\pm 0.06\mathrm{J}/{\mathrm{cm}}^2$, Table 2).

The ablation of cortical bone with femtosecond laser radiation leads to the formation of irregular and ragged surfaces, with no traces of melting or cracking. The diffraction peaks of hydroxyapatite in untreated cortical bone are much wider than those of synthetic crystallized hydroxyapatite, reflecting the partially amorphous nature of this mineral in bone and the extremely small size of the crystals (platelets with a diameter of 10 to 20 nm and less than 1 nm thick⁵³). The width of most peaks remains unchanged by the laser treatment, but a prominent $(\begin{array}{ccc}3& 0& 0\end{array})$ narrow peak appears at 32.8 deg, showing that the laser treatment led to the recrystallization of hydroxyapatite, forming larger crystals with the $(\begin{array}{ccc}1& 0& 0\end{array})$ prismatic plane predominantly parallel to the sample surface. This alteration affects only a small volume of material, because no other changes are observed in the diffractograms. This is corroborated by the fact that the $\mathrm{Ca}/\mathrm{P}$ ratio, determined by XPS analysis, is not affected by the laser treatment and remains similar to the theoretical value for stoichiometric hydroxyapatite. No significant changes were observed for the hydroxyapatite by FTIR and micro-Raman.

Analysis of the FTIR, micro-Raman, and XPS results shows that the laser treatment leads to a reduction of the proportion of organic compounds in the surface layer of the material. In the FTIR spectra, the relative amplitude of several bands is associated with collagen decreases, which also occurs in the aromatic $\mathrm{C}─\mathrm{H}$ vibration bands in the Raman spectra. This may indicate that collagen is either partially denatured by the laser treatment (collagen can be denatured by exposition at temperatures as low as 40°C)^54,55 or partially eliminated from the surface due to the high temperatures attained during the ablative process. The decrease of the ${\mathrm{O}}_{531.1}/\mathrm{P}$ ratio observed by XPS analysis supports the latter hypothesis. The contribution of the carbonyl group oxygen to the ${\mathrm{O}}_{531.1}$ peak decreases with the laser treatment, and the ${\mathrm{O}}_{531.1}/\mathrm{P}$ ratio tends to the stoichiometric ratio in the phosphate group (${\mathrm{O}}_{531.1}/\mathrm{P}=4$). Similarly, the amount of carbon decreases (Table 1), also corroborating this interpretation. The disappearance of organic compounds from the most superficial layer of material in the laser-treated samples should also be accentuated by the redeposition of ablation debris, which must not contain carbon compounds, due to the high temperatures reached in the plume.

The XPS analysis shows that the $\mathrm{Ca}/\mathrm{P}$ ratio at the sample surface is different from the stoichiometric value both in the untreated and treated samples, but the difference decreases with the laser treatment. This variation may be explained by the existence prior to the laser treatment of an organic overlayer at the sample surface, whose thickness is reduced by this treatment, as suggested by the carbon and nitrogen concentrations in Table 1. Since the kinetic energy of P 2p photoelectrons is higher than the kinetic energy of Ca 2p electrons, the Ca 2p photoelectrons are more attenuated than the P 2p ones by this organic overlayer. A reduction of its thickness due to the laser treatment would explain the observed variation. On the other hand, in natural hydroxyapatite, Ca can be replaced by Mg in some crystallographic positions.⁵⁶ If Mg is quantified, the ratio $(\mathrm{Mg}+\mathrm{Ca})/\mathrm{P}$ is similar to the stoichiometric one (1.6666) within the limits of the experimental error.

The present results can be compared with those achieved by precedent authors using micro- and nanosecond pulse duration lasers. Jovanovic et al.⁵⁷ observed surface carbonization and melting in bone samples ablated with a Ho:YAG laser ($\lambda =2.1\mu \text{m}$, pulse duration 500 μs) using fluences from 90 to $130\mathrm{J}/{\mathrm{cm}}^2$. This effect is less pronounced but still present when an Er:YSGG ($\lambda =2.78\mu \text{m}$, pulse duration 500 μs, fluence $55\mathrm{J}/{\mathrm{cm}}^2$) is used. This important thermal effect can be expected due to the large pulse duration and fluences used. Modern Er lasers do not present problems to the same extent.¹⁹

In a previous paper, the authors concluded that the ablation of dentin, using experimental conditions similar to the present ones, occurred by a combination of photothermal and electrostatic mechanisms.²⁵ The present results indicate that the same mechanism occurs for bone. According to Rode et al.,⁵⁸ subpicosecond laser ablation has the advantage that the heat conduction time and the electron-to-ion energy transfer time are longer than the pulse duration. If the pulse energy exceeds the energy required to ionize the target surface, electrons will be removed from the sample, leaving behind the cations and creating an electric field due to charge separation. When the energy of the escaping electrons is higher than the remaining material-binding energy, material will be athermally ablated by a mechanism that the authors designated by electrostatic ablation. Electrostatic ablation occurs for radiation intensities higher than the electrostatic ablation threshold (${10}^{13}$ to ${10}^{14}\mathrm{W}/{\mathrm{cm}}^2$). If the absorbed power is lower than this threshold, the energy of the electrons is transferred to the lattice by electron-phonon interactions, and ablation occurs by a thermal mechanism, similar to the mechanism observed for nanosecond duration pulses and at a timescale that is orders of magnitude larger than the pulse duration. In actual practice, the spatial and temporal intensity distributions in the laser pulse result in a mixed interaction mode. In the present work, the intensities used were of the order of ${10}^{12}\mathrm{W}/{\mathrm{cm}}^2$, so a partially thermal ablation mechanism is to be expected. This may explain the thermal effects observed. Despite this thermal contribution to ablation, deterioration of the bone structure and of the underlying material is negligible. Femtosecond laser ablation allows an accurate and clear removal of bone without any traces of surface melting, carbonization, and cracking, putting this type of laser in clear advantage as compared to Ho:YAG and Er lasers for clinical applications.

5.

Conclusions

Laser ablation of cortical bone using an ultrafast laser with 500 fs pulse duration and 1030 nm wavelength and fluences higher than the ablation threshold allows controlled removal of bone with no evidence of melting, carbonization, or cracking. The ablated surfaces are flat, with an irregular and ragged appearance. The bone structure is preserved, and the surfaces are free of ablation debris. For the pulse fluences used (0.55 to $2.24\mathrm{J}/{\mathrm{cm}}^2$), ablation occurs by a combination of thermal and electrostatic mechanisms, the first mechanism predominating for lower fluences and the second one for higher fluences. This thermal mechanism leads to partial denaturation of collagen at the outermost layer of the remaining tissue. GI-XRD analysis suggests that some recrystallization of hydroxyapatite may also occur.